Hierarchically Porous Li4Ti5O12 Anode Materials for Li- and Na-Ion Batteries: Effects of Nanoarchitectural Design and Temperature Dependence of the Rate Capability



Integrated design of both porous structure and crystalline morphology is expected to open up the way to a new class of materials. This report demonstrates new nanostructured Li4Ti5O12 materials with hierarchically porous structures and flower-like morphologies. Electrochemical studies of the electrodes of Li-ion and Na-ion batteries clearly reveal the advantage of nanoarchitectural design of active materials. In addition, the temperature dependence of Na+-insertion/extraction capacity in relation to Li4Ti5O12 electrodes is for the first time evaluated and it is found that elevation of the cell operating temperature effectively improves the rate capability of the Na-ion batteries. Based on the new findings, it is suggested that specially designed Li4Ti5O12 materials allow for high-performance Na-ion batteries that are available as large-scale storage devices for applications such as automotive and stationary energy storage.

1 Introduction

As growing worldwide concerns about energy storage for the intermittent renewable energy sources, such as wind, solar and wave, in substitution for fossil fuels and nuclear energy, secondary Li-ion batteries have been receiving intense interest over the last decade.[1] The issue of Li supplies (for instance, the upsurge in the cost) is yet to be an obstacle for the extension to large-scale storages. In order to implement the demands for massive energy storage, recent efforts have focused on the development of new types of rechargeable batteries. Among them, Na-ion batteries exhibit significant potential as an alternative storage device[2, 3] due to inexhaustible Na resources around the world and thereby low cost, and the predicted comparable energy density with Li-ion batteries based on the Na+/Na redox potential of –2.71 V (vs. SHE). Besides, the lower Lewis acidity of Na+ than that of Li+ leads to the rapid interfacial Na+ transport between electrolyte and host materials because of the lower activation energy for desolvation of Na+ ions.[4] Since the main limiting factor in the rate capability of Li-ion batteries is the activation barrier of Li+ transfer at a solid/electrolyte interface, Na-ion batteries have greater potential to deliver high-rate performance. Recently, Nose et al. reported that the mixed phosphate compound, Na4Co3(PO4)2P2O7, and its substitution products show both high potential and high rate performance as a cathode for Na-ion batteries,[5] which allows to materialize commercially applicable Na-ion batteries if a high-performance anode material with sufficient rate capability will be developed.

To date, the most intensively studied anode materials for Na-ion batteries are carbon-based materials.[6] Although graphite, which is the most common anode for Li-ion batteries, is unavailable for a Na+-insertion host,[7] disordered carbons still remain one of the candidates for achieving high capacity. However, as is the case with the carbon-based negative electrode for Li-ion batteries, several issues have been pointed out regarding safety and cyclability.[3] Most notably, the operating voltage close to the electroplating potential of Na raises concern about a Na-metal dendrite formation, which triggers internal short, especially at high rates.[8] The higher reactivity of metallic Na than Li make this issue more problematic, indicating that the use of anode materials with the low operating potential is undesirable for practical uses. As alternative candidates for carbon electrodes, titanium-based compounds like, for instance, TiO2 (amorphous,[9] anatase[10] and bronze-phase[11], Na2Ti3O7,[12] Na4Ti5O12,[13] Na2Ti6O13,[14] NaTi3O6(OH)·2H2O[15] and Na0.66(Li0.22Ti0.78)O2[16] have been intensively studied. These candidates, however, neither completely dispel the concern on Na metal electroplating due to the low storage voltage, nor satisfy capacity and rate-performance required for commercially available batteries.

Very recently, the availability of spinel lithium titanate (Li4Ti5O12) as an anode for Na-ion batteries has been disclosed by Sun et al.[17] As is well known, Li4Ti5O12 is one of the most promising anode materials for Li-ion batteries due to its unique characteristics, such as “zero-strain” in the lattice on charge/discharge and the higher Li+-insertion/extraction potential (≈1.55 V vs. Li+/Li) with the theoretical capacity of 175 mAh g−1,[18] resulting in excellent cycle-life simultaneously with enough safety. In addition, the inherently poor electric conductivity (10−13–10−8 S cm−1) of Li4Ti5O12[19] can be resolved by forming Li4+δTi5O12 phase with high enough conductivity for battery applications,[20] which fulfills the requirement for a high-rate electrode. The avoidance of laborious carbon-coating is also advantageous for low-cost and high power density.[21] With respect to the application of Li4Ti5O12 to Na-ion batteries, Sun et al. reported the Na+-insertion/extraction potential of ≈0.9 V (vs. Na+/Na) and theoretical capacity of 175 mAh g−1, according to the following three-phase separation mechanism.[17]

display math(1)

In this regard, the Li4Ti5O12 electrode demonstrates significant potential as a high-performance anode for Na ion batteries. Nevertheless, significant challenges for the high-rate anode still remain,[17, 22] because of the remarkably slow Na+ diffusion coefficient of Li4Ti5O12 (≈10−16 cm2 s−1)[23] compared to the Li+ diffusion coefficient (10−13–10−9 cm2 s−1).[24]

Meanwhile, recent efforts have focused on designing porous structures[25, 26] and nanoarchitectures, such as nanoparticles,[27] nanowires,[28] and nanosheets,[29] for enhancing the electrochemical performance of Li4Ti5O12 anodes for Li-ion batteries. The formulation of hierarchical porous structure in active materials offers crucial benefits; small pores provide higher specific surface area, thereby more active surface, while large pores contribute to better penetration of electrolyte in the electrode and reduce diffusion pathways of ions in solid.[30] The nanostructured crystallites also facilitate higher charge/discharge rates with the aforementioned effects and deliver high rate durability.

Here, we synthesize a novel Li4Ti5O12 monolithic material with complex nanoarchitectures of hierarchically porous structure as well as flower-like morphologies on the macropore surface. The sol–gel process accompanied by spinodal decomposition yields interconnected porous structure with uniform through-pores,[31, 32] which is advantageous to rapid penetration of relatively viscous electrolyte comparing to the so-called three-dimensionally ordered macroporous (3DOM) structure with a small window size tailored by colloidal templates.[25] The mesopores in the macropore skeletons as interstices of nanocrystallites facilitate rapid charge transfer. In addition, the flower-like structured crystallites on the surface of the macropore skeletons, which are formed simply by a treatment in LiOH aq.,[33] increase active sites. In the experiments described herein, the electrochemical capabilities of designed Li4Ti5O12 electrodes for Li- and Na-ion batteries were explored in detail with three-electrode cell systems. Moreover, we demonstrate the correlation between cell temperature and Na-ion storage performance, especially rate-capability, which offers new insights for the progress in practical use of Na-ion batteries.

2 Results and Discussion

2.1 Synthesis and Characterization of Porous Li4Ti5O12 Monoliths

Monolithic TiO2 gels with tunable macropores have been prepared via the alkoxy-derived titania sol–gel system incorporating spinodal decomposition in our previous work.[32] The subsequent treatment of porous TiO2 gels in LiOH aq. under the relatively mild condition gives rise to platy layered hydrous lithium titaniate (LHLT)[34] crystallites on the macropore surface through dissolution-reprecipitation, leading to the flower-like structures.[33] The crack-free monolithic shape as well as the interconnected macroporous structures are preserved through calcination accompanied by crystal phase transition, as shown in Figure 1a,b. The flower-like structures composed of platy crystallites with the thickness of ≈30 nm can be retained up to 700 °C (Figure 1c,e). When the sample was calcined at 800 °C, the platy crystallites on the macropore skeletons transformed into the rod-like ones due to the sintering densification (Figure 1d,f). The insets of Figure 1e,f present the cross-sections of macropore skeletons, implying that it consists of granular crystallites with the size of 30–80 nm when calcined at 700 °C, while the sample calcined at 800 °C displays the smooth cross-section.

Figure 1.

a) Appearance of the Li4Ti5O12 monoliths calcined at 700 °C. SEM images of the macroporous structure of the samples calcined at b,c) 700 °C and d) 800 °C. Magnified images of the samples calcined at e) 700 ºC and f) 800 ºC; insets are the cross-section of the macropore skeletons.

Figure 2a demonstrates the nitrogen physisorption isotherms and the corresponding mesopore size distributions (inset). The sample calcined at 700 °C possesses substantial amounts of mesopores larger than 10 nm, which are derived from the interstices of the crystallites constituting macropore skeletons. By contrast, almost no micro- and mesopores remain at 800 °C, which is consistent with the SEM images. The specific surface areas calculated by the BET method of the samples calcined at 700 °C and 800 °C are 30 m2 g−1 and <5 m2 g−1, respectively.

Figure 2.

a) Nitrogen adsorption-desorption isotherms and mesopore size distributions calculated by the BJH method (inset). b) XRD patterns of the samples calcined at different temperatures.

As confirmed by the XRD patterns in Figure 2b, the as-dried sample is composed of anatase (macropore skeletons) and LHLT (flower-like structures).[33] With increasing calcination temperature, the crystal phases transform to TiO2/LiTiO2, TiO2/Li2TiO3, and finally to the Li4Ti5O12 single phase. In our previous study,[33] the samples treated in LiOH aq. were completely washed with H2O to remove all the residual Li species, ending up with TiO2/Li4Ti5O12 composites after calcination. In this study, we carefully modified the treatment condition in LiOH aq. and washed the gels with EtOH to avoid excessive dissolution of Li from the treated samples, which allows for the development of the single-phase Li4Ti5O12 spinel framework (see Supporting Information Figure S1).

Transmission electron microscopy (TEM) observation of the sample calcined at 700 °C was conducted for investigation of the crystalline features of the flower-like structure. The TEM images (Figure 3a,b) and the corresponding selected-area electron diffraction (SAED) patterns shown in Figure 3c reveal that each petal of the flower-like structure consists of Li4Ti5O12 single-crystalline nanosheets of 100–200 nm in size. A representative high-resolution TEM (HR-TEM) image taken from the nanosheet (Figure 3d) clearly shows three lattice fringes with spacings of 0.41 nm, 0.48 nm and 0.49 nm, which correspond to the d spacings of (002), (11inline image) and (111) planes of Li4Ti5O12, respectively. It is therefore confirmed that the observation is along the [inline image10] direction. A TEM observation of the sample calcined at 800 ºC implies that each rod-like crystallite on the macropore surface is also composed of a Li4Ti5O12 single crystal (Supporting Information Figure S2).

Figure 3.

a,b) TEM images of the sample calcined at 700 °C. c) The electron diffraction pattern of the single nanosheet illustrated in (b). d) HR-TEM image of the nanosheet shown in (b). The inset shows a crystal structure of spinel lithium titanate viewed along [inline image10] zone axis.

2.2 Electrochemical Properties of Porous Li4Ti5O12 in Li-Ion Batteries

For exploring the benefits granted by the nanostructure, the capabilities of the obtained porous Li4Ti5O12 materials for Li-ion battery electrodes were investigated in a half cell. Herein, Li-extraction is expressed as discharge, regarding Li4Ti5O12 as an anode. The charge-discharge profiles and the rate properties of the samples are presented in Figure 4a–c. The samples calcined at 700 °C and 800 °C exhibit flat plateaus at 1.55 V (vs. Li+/Li), indicating the characteristic of two-phase reaction of Li4Ti5O12/Li7Ti5O12,[18] and the discharge capacities of 165 mAh g−1 and 150 mAh g−1, respectively, at 0.1C. It is worth noting that the hierarchically porous Li4Ti5O12 with flower-like structures (the sample calcined at 700 °C) delivers the reversible high rate capacities of 155 mAh g−1 and 121 mAh g−1 at 10C and 30C, respectively, even without carbon-coating due to the substantial increase in electric conductivity by Li+ insertion to Li4Ti5O12.[20, 21] On the other hand, the capacity of the sample calcined at 800 °C, which can be identified as relatively large crystallites (from sub-micrometer to micrometer-sized) with negligible amount of micro-/mesopores and thereby less nanostructural effects, decreases more steeply as increasing the C-rate. In light of these results, the enhanced rate property of the sample calcined at 700 °C can be attributed to the pores in the macropore skeletons and the flower-like structures, in addition to the smaller crystallite size due to the lower calcination temperature. The hierarchically porous Li4Ti5O12 with flower-like structures also exhibits high durability, as shown in Figure 4d. After 500 cycles at 5C, more than 98% of the initial capacity was still retained, indicating that the tailored structures are stable during the electrochemical Li+ insertion/extraction because of the zero-strain topotactic transformation of Li4Ti5O12/Li7Ti5O12.[18]

Figure 4.

Electrochemical properties of porous Li4Ti5O12 electrodes for Li-ion batteries. Galvanostatic charge and discharge curves (3rd cycle) of the samples calcined at a) 700 °C and b) 800 °C at different C-rates. c) Rate properties of the samples calcined at 700 °C and 800 °C. The closed and open symbols represent charge (Li+ insertion) and discharge (Li-extraction) capacities, respectively. d) Cycle performance of the nanostructured Li4Ti5O12 electrode calcined at 700 °C at a rate of 5C. Cycling test was carried out after the rate property measurement of (c).

In order to examine in more detail the effect of nanostructure on electrode capability, electrochemical impedance spectroscopy (EIS) measurements were carried out. Figure 5a,b show the variation in the Nyquist plots with decreasing the potential from 1.60 V to 1.45 V (vs. Li+/Li). In cases of both electrodes, the onset potential for Li+ insertion is around 1.60 V and the cathodic peak centered at around 1.53 V is observed in the cyclic voltammetry (CV) curves (Figure 5c). The Nyquist plots form a small semicircle related to the high frequencies and a linear component related to the low frequencies. It is worth noting in relation to the difference between two electrodes that the sample calcined at 700 °C exhibits very steep sloping line indicating finite diffusion by virtue of the nanostructure,[30] whereas the deviation of the sloping line from the Warburg impedance is small (thereby semi-infinite diffusion) in the case of the sample calcined at 800 °C. Accordingly, the nanostructured electrode both with hierarchically porous structures and flower-like structures achieves high rate capability, as mentioned above.

Figure 5.

Nyquist plots of the electrodes calcined at a) 700 °C and b) 800 °C in 1 M LiClO4 (EC/DEC) electrolyte measured at different potentials. c) Cyclic voltammogram at a scan rate of 0.1 mV s−1.

2.3 Temperature Dependence on Electrochemical Performance of Porous Li4Ti5O12 in Na-Ion Batteries

Unlike the topotactic transformation of Li4Ti5O12 during Li+ insertion/extraction, the charge and discharge processes involve a lattice volume change (≈13%) in the Li4Ti5O12/LiNa6Ti5O12 phase transition in the case of Na-ion batteries.[17] Moreover, the slow Na+ ion diffusion rate in the spinel host also raises the requirement for the nanostructural design of Li4Ti5O12 electrode.[23] Figure 6a displays the charge and discharge curves of the nanostructured electrode calcined at 700 °C at 0.1C. The 1st charge (Na+ insertion) curve showed the plateau at –2.65 V (vs. Ag+/Ag, corresponding to 0.77 V vs. Na+/Na) and the plateau shifted to the positive direction from the 2nd cycle, while the discharge curves exhibited the similar shape, which is consistent with the previous reports.[17, 22] It is deduced that minor structural rearrangement might take place during the first charging process. The discharge curve contains three different regions, which are slope region (< –2.50 V (vs. Ag+/Ag)) and two plateaus at –2.45 V and –2.24 V (vs. Ag+/Ag). This is also supported by the CV profile in Figure 6b; three anodic peaks at around –2.57 V, –2.42 V and –2.24 V (vs. Ag+/Ag) can be observed. It suggests that the Na-extraction process is not a reverse reaction of Na+ insertion but a more complicated pathway.[23] Although the discharge capacity at 0.1C reaches 158 mAh g−1, the rate-performance undoubtedly declined compared to the case of Li+ insertion/extraction, as shown in Figure 6c,d. The discharge capacities at 1C and 2C were 104 mAh g−1 and 82 mAh g−1, respectively.

Figure 6.

Electrochemical properties of the Li4Ti5O12 electrode calcined at 700 °C for Na-ion storage. a) Galvanostatic charge and discharge curves (1st to 5th) at 0.1C. b) CV curves at a scan rate of 0.1 mV s−1. c) Charge-discharge curves (3rd cycle) at different C-rates. d) Cycle performance at different current rates.

The Nyquist plots of the nanostructured Li4Ti5O12 electrode calcined at 700 °C in the electrolyte for Na-ion batteries are illustrated in Figure 7. The size of semicircle in high frequency region is similar to those for Li-ion batteries (Figure 5a). This is because the Li7Ti5O12 phase with good conductivity is formed during Na+ insertion like the case during Li+ insertion.[20] In addition, it also implicates that the resistivity of Na+ ion transport at the solid/electrolyte interface is small in the case of Na-ion battery system due to the lower Na+-solvent interaction (compared to Li+) derived from the lower Lewis acidity.[4] We therefore speculated that the major reason of the poorer rate-performance is sluggish Na+ ion diffusion rate and crystal phase migration kinetics of LiNa6Ti5O12/Li4Ti5O12/Li7Ti5O12.[23] In fact, the Nyquist plots show the sloping line with the angle of about 45°, which means semi-infinite diffusion of Na+ ion in the active materials.

Figure 7.

Nyquist plots of the electrode calcined at 700 °C in 1 M NaClO4 (EC/DEC) electrolyte measured at different potentials.

The temperature dependency on the rate performance was therefore examined, as demonstrated in Figure 8a–c. Obviously, the rate capability was effectively improved with elevating temperature. The discharge capacities at 1C and 2C were amended up to 127 mAh g−1 and 113 mAh g−1, respectively, when the temperature was raised from room temperature to 40 °C. At 60 °C, the rate performance was further enhanced amounting to 105 mAh g−1 and 63 mAh g−1 at 10C and 30C, respectively, while the significant degradation of the electrode through cycling was observed together with impaired coulombic efficiency. More surprisingly, the elevation of operating temperature has a considerable influence even on the sample calcined at 800 °C, which delivers negligible capacity at room temperature (see Supporting Information Figure S3). The discharge capacity was improved up to 100 mAh g−1 at 0.1C on raising temperature to 60 °C. The improvement in the electrochemical performance with increasing temperature can be evidently observed in the CV profiles (Supporting Information Figure S4) as well. In the case of the sample calcined at 800 °C, the redox couple was ambiguous at room temperature. On the other hand, the well-defined cathodic and anodic peaks can be detected in the CV curves performed at 60 °C. These results also strongly suggest that the increase in operating temperature is a critical matter to achieve a high-performance Na-ion battery due to the accelarated Na+ ion diffusion rate and movement of crystal phase boundaries of LiNa6Ti5O12/Li4Ti5O12/Li7Ti5O12.

Figure 8.

Comparison of charge and discharge curves of the nanostructured Li4Ti5O12 electrode calcined at 700 °C in 1 M NaClO4 (EC/DEC) electrolyte at different rates measured at different operating temperatures; a) 40 °C and b) 60 °C. c) Correlation between cell temperature and capacity at various rates. The closed and open symbols represent charge (Na+ insertion) and discharge (Na+ extraction) capacities, respectively. d) Cycle properties at 1C conducted at different operating temperatures.

Figure 8d presents the cycle performance of the nanostructured electrode at 1C, which compares the durabilities of the electrode at room temperature and 40 °C. The discharge capacity of the 100th cycle was 93% of the initial capacity when the cycle tests were conducted at room temperature. Since the insertion and extraction of Na+ ion into Li4Ti5O12 involve volume change, the cycle performance declined compared to the case for Li-ion batteries. In the meantime, the capacity retention at 40 °C after 100 cycles was only 70%. This may be attributed to degradation of the electrode caused by the decomposition of the electrolyte and the detachment of the active materials from the electrode due to the deterioration of the binder. The implication of these results is that the optimization of electrode-preparation process including a selection of binders and improvement of the electrolyte for high-temperature operation are still required for practical use.

3 Conclusion

Fabrication of the complex nanoarchitectures that combines the hierarchical porous structure and flower-like structure into a Li4Ti5O12 monolith has been achieved by the treatment of TiO2 gels in LiOH aq. under a mild condition and subsequent calcination. The sample calcined at 700 °C possessed well-defined mesopores as interstices of Li4Ti5O12 crystallites in the macropore skeleton simultaneously with preserving the established flower-like structures formed on the macropore skeletons. The flower-like structure on the macropore surface is advantageous to increase the efficiency of solid/electrolyte contact. On the other hand, only a few mesopores remained and the flower-like structures transformed to more densely aggregated structure with rod-like crystallites, when the calcination temperature was increased to 800 °C.

The nanostructured Li4Ti5O12 electrodes (calcined at 700 °C) exhibited remarkably high rate performance of 146 mAh g−1 and 105 mAh g−1 at 10C and 30C even without carbon-coating due to the well-designed nanoarchitecture, which is confirmed by the comparison with the less porous sample (calcined at 800 °C). The sample calcined at 700 °C also worked well when applied to an anode for Na-ion batteries, whilst it showed the poorer rate capability due to the slow diffusion kinetics of Na+ ion. It was, however, found that the increase in the operating temperature drastically improves the rate capability. The new findings provided in this manuscript suggest that a Li4Ti5O12 electrode has a potential to offer a viable Na-ion battery with enough rate performance for practical uses. It is required that the development of an electrolyte as well as a binder that are safe and durable for use at 40–60 °C.

4 Experimental Section

Chemicals: Titanium (IV) n-propoxide (Ti(OPr)4) and poly(ethylene glycol) (PEG, Mv = 10 000) were purchased from Sigma-Aldrich Co. Ethyl acetylacetonate (EtAcAc) and 1-propanol (PrOH) were purchased from Tokyo Chemical Industry Co., Ltd. Lithium hydroxide monohydrate (LiOH·H2O) and ammonium nitrate (NH4NO3) were obtained from Kishida Chemical Co., Ltd. All reagents were used as received. Distilled water was used in all experiments.

Synthesis of Porous Li4Ti5O12 Monoliths: Porous TiO2 monoliths were prepared as reported previously.[32] In a typical synthesis, 10 mL of Ti(OPr)4, 7.0 mL of PrOH, and 5.0 mL of EtAcAc were mixed in a glass tube. After obtaining a homogeneous yellow solution, 0.850 g of PEG was added followed by stirring at 60 °C until PEG was completely dissolved. The solution was then cooled to 40 °C, and 2.0 mL of 1.0 M NH4NO3 aq. was added slowly with vigorous stirring. After mixing for 3 min, the obtained homogeneous solution was kept at 40 °C for 24 h. The obtained gel was subjected to the sequential solvent exchange from EtOH to H2O. Thus prepared TiO2 wet gel (about 0.55 g after drying) was treated in 25 mL of 0.5 M LiOH aq. at 100 °C for 24 h in a Teflon-lined autoclave. After washing with ethanol (EtOH) at 60 °C for 2 h five times, the gel was dried at 60 °C. The gel was then calcined at different temperatures for 2 h.

Characterization: The microstructures of the fractured surfaces of the samples were observed using SEM (JSM-6060S, JEOL), field emission scanning electron microscopy (FE-SEM) (JSM-6700F, JEOL), and TEM (JEM-2200FS, JEOL, equipped with a CEOS image corrector). Nitrogen adsorption–desorption apparatus (Belsorp mini II, Bel Japan Inc.) was employed to characterize the meso- and micropores of the samples. The samples were degassed at 200 °C under vacuum for more than 6 h. The crystal structure was confirmed by powder XRD (RINT Ultima III, Rigaku Corp., Japan) using Cu Kα (λ = 0.154 nm) as an incident beam.

Electrochemical Measurements: All the electrochemical tests were carried out on a VMP3 potentiostat (Bio Logic Science Instruments) at room temperature (≈22 °C) unless otherwise noted. The Li4Ti5O12 electrode slurries comprised 85% active material, 10% acetylene black, and 5% carboxymethylcellulose sodium (CMC-Na, Kishida Chemical Co., Ltd.) binder, in distilled water. The electrode slurry was coated onto an Al foil. The electrode was then dried under vacuum at 60 °C for 24 h. The typical density of the active material was approximately 0.8–1.0 mg cm−2. For the electrochemical tests for Li-ion batteries, the electrodes were tested in a three-electrode cell consisting of a 1 m LiClO4 electrolyte in 1:1 (v/v) ethylene carbonate (EC)/diethyl carbonate (DEC) (Kishida Chemical Co., Ltd.) and Li metal as both the counter and reference electrodes. Cyclic voltammetry (CV) measurements and constant current charge-discharge tests were carried out with 1.0 and 2.0 V (vs. Li+/Li) cutoff potentials at 0.1 mV s−1. In the case of the tests for Na-ion batteries, the electrolyte was a 1 M NaClO4 (Sigma-Aldrich Co., LLC.) in EC/DEC (1:1 in v/v) and Na metal was used as the counter electrode. The reference electrode was prepared with Ag wire immersed in the electrolyte containing 50 mM AgCF3SO3 (Kishida Chemical Co., Ltd.). The potential of the obtained silver reference electrode was recorded as ≈3.42 V (vs. Na+/Na). The cells were thermostated and cycled between fixed potential values from –2.90 to –1.90 V (vs. Ag+/Ag) in the CV and charge-discharge tests. Electrochemical impedance spectroscopy (EIS) measurements were conducted in the 10 mHz to 100 kHz frequency range. For consistency, the currents were calculated only on the active material.


The present work was supported by the Grant-in-Aid for JSPS Fellows (No. 24·31 for G.H.) from Japan Society for the Promotion of Science (JSPS).